In this paper, we continue to study the geodetic stations adjusted coordinates’ statistical properties in permanently operated networks. Analytical reconstructing the variance-covariance matrix of a real vector solution obtained from processing a redundant set of the same GNSS observation session’s measured vectors is performed. The mentioned matrix is compared with the same one, of the vector solution, resulted on the measured lines’ total set basis. Common properties of both arrays are revealed, as well as discrepancies in features for different types of solutions. As a suitable tool for studying the latter, the corresponding absolute way is offered: the values of the vector network points’ spatial coordinates and their common covariance matrix of errors in the proper reference frame of the solution. It is shown that in a method built on the basis of an incomplete redundant set of measured vectors (compared to the total set), the probability of local deformations i. e. common displacements of the stations compact groups increases. It is noted that they can be interpreted as migrations of the reference frame’s origin reproduced by this vector network in its territory. It is shown which parameters should be used to assess the probable level of these deformations within the area covered by the vector network